*Official Benchmark Thread* - Post it here or it didn't happen :D

[Talon]

Blown away by the performance uplift of the 4090 over 3090/Ti. I'm headed to MC in the morning to try my hand at grabbing one. Sounds like they're doing some sort of lottery and I'm sure I'll have to compete against freaking scalpers. Last time we had to line up, so it forced scalpers to wait out there with us, this time they can show up at open for a chance to resell or return that card, so nothing invested. Total BS and rubs me wrong TBH.

[Etern4l]

20 minutes ago, Papusan said:

Why? Because they went the Dell route and prefer less noise over offering max cooling performance? 4090 use less power than previous 3090 Ti in benches/games and still run hotter. This with less power consumption. Here there should be some improvements from the AIB partners air cooled cards. Yep, not all run their cards fans at max rpm, but why cap the fan's maximum cooling performance.

 

All this reminds me about how POW castrated their GTX 580 cards cooling capacity with an added 75% fan rpm cap in the vbios. A stupid move if boost bins determines out from max gpu temps. Aka lower temps will provide higher boost clocks.

 

To finish up my testing I of course had to check out the cooling performance. To do this I ran two different tests. I used AIDA64’s Stress Test run for a half-hour each to warm things up. Then I documented what temperature the GPU leveled out at with the stock fan profile and then again with the fans cranked up to 100%. With the stock profile, the Nvidia RTX 4090 Founders Edition did better than I expected it to do given its power levels. I leveled off at 77c which is up in the top ¼ of our charts but is lower than cards like the GTX 1080 and 1080Ti did and just one degree above the RTX 3080. Then with the fans cranked up, the RTX 4090 Founders Edition cooled down but not nearly as much as any of the other cards I’ve tested in the past with it running 67c at full fan speed on both fans. This shows that while the cooler did great with the stock profile, there isn’t too much headroom left in it. I would say that Nvidia’s fan profile is spot on to offer good cooling while keeping the noise down relatively. Of course, if they ran the fans at similar speeds to past Founders Edition cards there would be more headroom left. The RTX 3090 Ti which is a similarly sized card was running cooler at 73c under load and 61c at 100% fan speed which means it had a little more headroom at 12c. 

 

https://lanoc.org/review/video-cards/8600-nvidia-rtx-4090-founders-edition?start=9

 

Does Nvidia lock out software fan control on those FE cards? Normally I would almost immediately customise the fan curve using Afterburner...

[Papusan]

21 minutes ago, Etern4l said:

 

Does Nvidia lock out software fan control on those FE cards? Normally I would almost immediately customise the fan curve using Afterburner...

Nope. But you are stuck with the added max fan speed. 

 

I remember when Dell screwed us who bought the AW17. You had to cut one of the fan cables to get the max rated fan speed (max fan speed was only possible when you flashed new bios). And the fan profile in bios was heavly screwed (the fan didn't kick in before the Cpu reached throttle temp and clock speed dropped). Never fixed. The Dell way.

[Etern4l]

6 minutes ago, Papusan said:

Nope. But you are stuck with the added max fan speed. 

 

Ah got it, so they went with larger fans, but lower RPM to keep the noise down. I guess that kind of makes sense, I mean those FE cards where targetting heavy OCing. On the other hand, good luck during a 40C summer heatwave! It's just that the resulting height + 16 pin connector adapter are so annoying, needs a fairly wide case with no side fans.

[electrosoft]

48 minutes ago, Talon said:

Blown away by the performance uplift of the 4090 over 3090/Ti. I'm headed to MC in the morning to try my hand at grabbing one. Sounds like they're doing some sort of lottery and I'm sure I'll have to compete against freaking scalpers. Last time we had to line up, so it forced scalpers to wait out there with us, this time they can show up at open for a chance to resell or return that card, so nothing invested. Total BS and rubs me wrong TBH.

 

Good luck bro and yeah the performance uplift is substantial. Just on paper you can see with the die shrink, ~5500+ more cores and +600-700 on the boost clocks. Looking forward to your results. The RT uplift looks substantial too looking at HWU testing.

 

I think (hope?) Scalpers are in for a rude awakening this time around but you never know.

 

Are you targeting a specific card or will you take whatever you can get?

 

[Talon]

2 minutes ago, electrosoft said:

 

Good luck bro and yeah the performance uplift is substantial. Just on paper you can see with the die shrink, ~5500+ more cores and +600-700 on the boost clocks. Looking forward to your results. The RT uplift looks substantial too looking at HWU testing.

 

I think (hope?) Scalpers are in for a rude awakening this time around but you never know.

 

Are you targeting a specific card or will you take whatever you can get?

 

 

I want a Strix or MSI SuprimX even though I hate the way the Asus looks. Asus has been good to me with RMAs lately and they honor second hand warranty I think. MSI can be hit or miss, usually a huge miss from my experience dealing with them, but their graphics cards are usually built well. 

[tps3443]

1 hour ago, Talon said:

 

I want a Strix or MSI SuprimX even though I hate the way the Asus looks. Asus has been good to me with RMAs lately and they honor second hand warranty I think. MSI can be hit or miss, usually a huge miss from my experience dealing with them, but their graphics cards are usually built well. 

I think one of those massive 43” 4K OLED monitors, and a RTX4090 overclocked to the moon would be absolutely marvelous. 

[Mr. Fox]

3 hours ago, tps3443 said:

The RTX4090 is very fast. But we’ve almost reached a point where we don’t even need all that lol. It would be fun for sure. But, I play at 2560x1440P. I’m just not a fan of a 4K monitor yet. 
 

 

4090 is awesome. But not needed at all. It throws down at high res though!!! 

If you look at the minimum and recommended system requirements on most games, those without ray tracing run great with old hardware (old as in Maxwell and newer) that only those that want to crank up the ray tracing need a high end GeFarts GPU. Most people don't play at 4K. Most PC gamers are console jockeys that figured out there is a better way to live, or old-school gamers that never invited console demons into their hearts. Most of them don't buy high-end CPU, GPUs or 4K monitors, and they don't need ray tracing to have fun playing their games. And, they're happy holding a steady 60 FPS. It is only us in the elite extreme camp that care more about the weapons of mass destruction than the waging of war.

 

We all already know this. I was reminded of it when I tested a couple of games on this chintzy little Zotac 3060 and noticed no difference in gameplay quality than I do using the shunt-modded 2080 Ti monster and 3090 KPE, both of which mop the floor with it in benchmarks.

[Reciever]

39 minutes ago, tps3443 said:

I think one of those massive 43” 4K OLED monitors, and a RTX4090 overclocked to the moon would be absolutely marvelous. 

It would if those monitors werent riddled with issues and compromises

 

The Aorus one is the only one that came close last I saw and its using compression because it cant actually hit 4k120hz

 

Honestly though I was just looking for a 40ish" panel with 4k and freesync

 

My 42" TV still works but just looking ahead to see what I would get as I dont want the hassle of the newer SmartTV's and setting it up on its own VLAN to prevent it from scanning my network. 

[Mr. Fox]

22 hours ago, tps3443 said:

It looks like Newegg has officially sold out of the 13900K, and sold out of the 13900KF. 

They had more 13900K available for pre-order this afternoon when I successfully resisted temptation. It wasn't easy, but I somehow managed. To make it easier, I should bury my head in the sand and ignore the world around me. But, then I would have to change my name to Mr. Ostrich.

 

Right now, the foreknowledge that I would be required to stop using Windows 7 due to a total absence of driver support is enough to make the idea of not flushing a small fortune down the toilet on a 4090 not really all that difficult for me, but the CPU is always the star in my show. 

 

If Roman pops the lid on a 13900K and confirms there are no hidden accidents looking for a place to happen that is going to make common sense much harder for me to find.

[electrosoft]

58 minutes ago, Mr. Fox said:

They had more 13900K available for pre-order this afternoon when I successfully resisted temptation. It wasn't easy, but I somehow managed. To make it easier, I should bury my head in the sand and ignore the world around me. But, then I would have to change my name to Mr. Ostrich.

 

Right now, the foreknowledge that I would be required to stop using Windows 7 due to a total absence of driver support is enough to make the idea of not flushing a small fortune down the toilet on a 4090 not really all that difficult for me, but the CPU is always the star in my show. 

 

If Roman pops the lid on a 13900K and confirms there are no hidden accidents looking for a place to happen that is going to make common sense much harder for me to find.

 

So much new hardware this fall......the temptation is real to at least get one of something.

 

I fully expect the power of Intel to compel you and you'll pick up a 13900k 🙂

 

 

[tps3443]

10 minutes ago, electrosoft said:

 

So much new hardware this fall......the temptation is real to at least get one of something.

 

I fully expect the power of Intel to compel you and you'll pick up a 13900k 🙂

 

 

He’s gonna end up buying one for sure. I thought he’d buy a 4090 too. That thing is a beast. 
 

As for myself I really want to just buy any 4090 tomorrow, just to have one because even the entry level power is hugely fast regardless.  But then I know a AIB model is gonna appear that I really really want instead. So, I’m gonna wait until every thing calms down after launch and carefully select one. Maybe a week maybe even 3 months lol.

 

I hate when new parts come out, something inside of us makes us feel like we desperately need to buy it just because it’s new and FASTTTTT. LOL. But then reality kinda sets in, and we don’t really actually need it.. But we still want it.

 

Just like @Mr. Fox said, we need to bury our heads. Maybe live on an island with no internet 😁

[Mr. Fox]

38 minutes ago, electrosoft said:

So much new hardware this fall......the temptation is real to at least get one of something.

 

I fully expect the power of Intel to compel you and you'll pick up a 13900k 🙂

I hope you are wrong. I am using the excuse that we need Roman to show us a delid is as easy as it was for 12900K before I frivolously throw money their way. If it is not as easy to delid, resisting will be easier, because I am not sure I would be happy a hot CPU that cannot be tamed by the correction of that design defect. When it didn't help the 5950X was when I threw in the towel on that ultimately hopeless cause. There we other factors involved, but the delid not helping it run cool enough to overclock further was the final nail in the coffin for that abortion.

 

I am also using EKWB's pursuit of a bare die solution as an excuse to wait and see. I guess we will find out how effective those excuses turn out to be.  SuperCool's bare die block being junk was a big letdown. 12900K/KS are awesome, but they would have been so much more so had that turned out to be designed in a way that it worked as well as it did for 10900K.

16 minutes ago, tps3443 said:

He’s gonna end up buying one for sure. I thought he’d buy a 4090 too. That thing is a beast. 
 

As for myself I really want to just buy any 4090 tomorrow, just to have one because even the entry level power is hugely fast regardless.  But then I know a AIB model is gonna appear that I really really want instead. So, I’m gonna wait until every thing calms down after launch and carefully select one. Maybe a week maybe even 3 months lol.

 

I hate when new parts come out, something inside of us makes us feel like we desperately need to buy it just because it’s new and FASTTTTT. LOL. But then reality kinda sets in, and we don’t really actually need it.. But we still want it.

 

Just like @Mr. Fox said, we need to bury our heads. Maybe live on an island with no internet 😁

Resisting 4090 will be much easier for me because it has the "no Windows 7 driver" bull$hit Micro$lop Mafia manipulation strings attached with it. I am both glad and disappointed at the same time. Disappointed that the Green Goblin drank from the Micro$lop cesspool, but glad that it gives me the strongest reason to say no to a GPU upgrade that I can identify.

[Reciever]

16 minutes ago, Mr. Fox said:

I hope you are wrong. I am using the excuse that we need Roman to show us a delid is as easy as it was for 12900K before I frivolously throw money their way. If it is not as easy to delid, resisting will be easier, because I am not sure I would be happy a hot CPU that cannot be tamed by the correction of that design defect. When it didn't help the 5950X was when I threw in the towel on that ultimately hopeless cause. There we other factors involved, but the delid not helping it run cool enough to overclock further was the final nail in the coffin for that abortion.

 

I am also using EKWB's pursuit of a bare die solution as an excuse to wait and see. I guess we will find out how effective those excuses turn out to be.  SuperCool's bare die block being junk was a big letdown. 12900K/KS are awesome, but they would have been so much more so had that turned out to be designed in a way that it worked as well as it did for 10900K.

Not that it would help you save any money but isnt Intel HEDT still in the rumor mill?

[Mr. Fox]

3 minutes ago, Reciever said:

Not that it would help you save any money but isnt Intel HEDT still in the rumor mill?

Yes it is. 😀 Still waiting to see more revealed on this, especially post-release intel (pun intended) about how far it can be overclocked.

[Ashtrix]

6 hours ago, Papusan said:

Why? Because they went the Dell route and prefer less noise over offering max cooling performance? 4090 use less power than previous 3090 Ti in benches/games and still run hotter. This with less power consumption. Here there should be some improvements from the AIB partners air cooled cards. Yep, not all run their cards fans at max rpm, but why cap the fan's maximum cooling performance.

 

All this reminds me about how POW castrated their GTX 580 cards cooling capacity with an added 75% fan rpm cap in the vbios. A stupid move if boost bins determines out from max gpu temps. Aka lower temps will provide higher boost clocks.

 

To finish up my testing I of course had to check out the cooling performance. To do this I ran two different tests. I used AIDA64’s Stress Test run for a half-hour each to warm things up. Then I documented what temperature the GPU leveled out at with the stock fan profile and then again with the fans cranked up to 100%. With the stock profile, the Nvidia RTX 4090 Founders Edition did better than I expected it to do given its power levels. I leveled off at 77c which is up in the top ¼ of our charts but is lower than cards like the GTX 1080 and 1080Ti did and just one degree above the RTX 3080. Then with the fans cranked up, the RTX 4090 Founders Edition cooled down but not nearly as much as any of the other cards I’ve tested in the past with it running 67c at full fan speed on both fans. This shows that while the cooler did great with the stock profile, there isn’t too much headroom left in it. I would say that Nvidia’s fan profile is spot on to offer good cooling while keeping the noise down relatively. Of course, if they ran the fans at similar speeds to past Founders Edition cards there would be more headroom left. The RTX 3090 Ti which is a similarly sized card was running cooler at 73c under load and 61c at 100% fan speed which means it had a little more headroom at 12c. 

 

https://lanoc.org/review/video-cards/8600-nvidia-rtx-4090-founders-edition?start=9

 

TPU had pictures of the FE card, I did not watch that GN video on Nvidia's PR engine about how great the Nvidia design is yet.

 

 

 

 

 

But yeah as we can see the 4090FE is having a thicker design, a bit fatter VC HSF area vs 3090Ti/3090 chassis. And the length is also a bit reduced vs the 3090/Ti design. This is how Nvidia was probably able to squeeze the GPU inside the massive 3 Slot brick design otherwise this GPU is not going to fit in that form factor of Triple Slot and would go to 4 slot a.k.a 3.x+. I thought 3090 was a giant card which will sag the hell out of a mobo, I plugged it already and yes it sags on the right side and over the time like 5-10 years, it will sag even more without some kind of holder.  An anti-sag GPU holder is a MUST for all 3090+ class cards.

 

RTX3090Ti FE - 2.2KG

RTX4090 FE - 2.18KG

 

Yeah that loss of length probably they shaved 100gm off.

 

And this is Nvidia FE design which is the best for Anti-Sag because their chassis is screwed on the entire lateral PCIe slot metal brace both sides 3 sets of screws on each side total 6, plus the width of the brace is proper for 3 Slot form factor vs say Gigabyte - 2 Slot on a 3.7 slot 4000 series design, EVGA / MSI / ASUS / GB all use 2 Slot brace and lower number of screws and uneven screw position unlike Nvidia's FE design for all their Ampere RTX3000 cards.

 

No wonder EVGA quit the scene on how expensive this whole GPU cooling thing is going to become on top of 12VHPWR PCIe5.0 cable adapter BS which is actually having lowest point at 30 times cycle. the cards are expensive to make, the cooling is expensive, the MSRP is expensive, the high TDP is demanding higher cabling requirement, fighting against the company which makes the GPUs themselves on top.

 

Nvidia really kicked into high gear since Ampere with FE design. And with the massive 4090 ADA102 core, the damn thing is very high tech power consuming device no wonder the size of the HSF is holding up and not leaving much space anymore. I wonder how RTX5090 and up hold.

 

[tps3443]

20 minutes ago, Reciever said:

Not that it would help you save any money but isnt Intel HEDT still in the rumor mill?

 

Honestly, anything HEDT that is new and coming from Intel will be crazy expensive. 
 

I imagine they’ll be competing with AMD Threadripper TRX40 socket right? Which is also crazy expensive. 
 

I highly doubt it’ll be like the X299 days. 

[Etern4l]

9 hours ago, Mr. Fox said:

Yes it is. 😀 Still waiting to see more revealed on this, especially post-release intel (pun intended) about how far it can be overclocked.

 

9 hours ago, tps3443 said:

 

Honestly, anything HEDT that is new and coming from Intel will be crazy expensive. 
 

I imagine they’ll be competing with AMD Threadripper TRX40 socket right? Which is also crazy expensive. 
 

I highly doubt it’ll be like the X299 days. 

 

If we are talking about X299 platform replacement, nothing is anywhere near what would seem like close to a release, right? Intel is obviously prioritising the Xeon platform, and happy with the use of Xeons in workstations. At one point they will have to move that to DDR5, I would guess mid 2023 at the latest. I suspect "X99 HEDT" is a bit like Nvidia's Titan - they probably feel it would work across purposes with the Xeon platform.

 

Xeon is what's meant by HEDT these days:

 

https://www.notebookcheck.net/16-cores-32-threads-and-77-MB-of-combined-cache-Intel-Sapphire-Rapids-Fishhawk-Falls-HEDT-CPU-leaks.629930.0.html

 

I mean Intel is in trouble. The 10nm process is really limiting them. The consumer CPUs can barely compete with AMD, then there is the ARM architecture. One thing I would like to see next in their high end consumer offering is at least 2x the amount of cache. This can work wonders even for gaming, as AMD demonstrated.

 

L3 cache (shared):

12900K: 30MB

13900K: 36MB (a sad increase, on the bright side they doubled the L2 cache to 2MB/core)

7950x: 64MB ...

Fastest gaming CPU with max boost clock of.... 4.5GHz: 96MB

 

The cache matters even more for professional applications, so I really hope Intel catches up and fast, somehow, as we shouldn't be at the mercy of TSMC these days.

 

My 12900KS visibly bottlenecks on cache/memory, so I'm not sure Raptor Lake with the extra e-Cores would help all that much, as the additional cores would just thrash the limited L3 cache.

The Xeon platform is not the answer either, the low core variants (ca. 20 cores or less), have about the same if not less amount of cache.

[jaybee83]

9 hours ago, tps3443 said:

I think one of those massive 43” 4K OLED monitors, and a RTX4090 overclocked to the moon would be absolutely marvelous. 

 

...sounds awesome, only that overclocking the 4090 wont net much additional performance. review roundup indicates like 5% performance added for +30-35% more power 😅 on the other hand, Roman showed 5% performance loss at 40%(!!!) less power. completely insane. granted, this will vary depending on the game/application used, but it doesnt promise any crazy additional increases.

 

5 hours ago, tps3443 said:

He’s gonna end up buying one for sure. I thought he’d buy a 4090 too. That thing is a beast. 
 

As for myself I really want to just buy any 4090 tomorrow, just to have one because even the entry level power is hugely fast regardless.  But then I know a AIB model is gonna appear that I really really want instead. So, I’m gonna wait until every thing calms down after launch and carefully select one. Maybe a week maybe even 3 months lol.

 

I hate when new parts come out, something inside of us makes us feel like we desperately need to buy it just because it’s new and FASTTTTT. LOL. But then reality kinda sets in, and we don’t really actually need it.. But we still want it.

 

Just like @Mr. Fox said, we need to bury our heads. Maybe live on an island with no internet 😁

 

you very likely chose the perfect gpu gen to bide ur time and take in the reviews first, no matter where i look or read, the message is always the same: about 5x volume at launch available vs. the 30 series, vendors like MC have literal truck loads of 4090s in stock and expectations are that initial stock will last around a week or so. i can imagine that some skus might be sold out more quickly than that due to focused demand, i.e. the FE or the Strix (although why anyone would spend 500 bucks more on basically the same cooling and power limits is beyond me...)

 

3 hours ago, Etern4l said:

 

 

If we are talking about X99 platform replacement, nothing is anywhere near what would seem like close to a release, right? Intel is obviously prioritising the Xeon platform, and happy with the use of Xeons in workstations. At one point they will have to move that to DDR5, I would guess mid 2022 at the latest. I suspect "X99 HEDT" is a bit like Nvidia's Titan - they probably feel it would work across purposes with the Xeon platform.

 

Xeon is what's meant by HEDT these days:

 

https://www.notebookcheck.net/16-cores-32-threads-and-77-MB-of-combined-cache-Intel-Sapphire-Rapids-Fishhawk-Falls-HEDT-CPU-leaks.629930.0.html

 

I mean Intel is in trouble. The 10nm process is really limiting them. The consumer CPUs can barely compete with AMD, then there is the ARM architecture. One thing I would like to see next in their high end consumer offering is at least 2x the amount of cache. This can work wonders even for gaming, as AMD demonstrated.

 

L3 cache (shared):

12900K: 30MB

13900K: 36MB (a sad increase, on the bright side they doubled the L2 cache to 2MB/core)

7950x: 64MB ...

Fastest gaming CPU with max boost clock of.... 4.5GHz: 96MB

 

The cache matters even more for professional applications, so I really hope Intel catches up and fast, somehow, as we shouldn't be at the mercy of TSMC these days.

 

My 12900KS visibly bottlenecks on cache/memory, so I'm not sure Raptor Lake with the extra e-Cores would help all that much, as the additional cores would just thrash the limited L3 cache.

The Xeon platform is not the answer either, the low core variants (ca. 20 cores or less), have about the same if not less amount of cache.

 

uve hit the nail on the head with this one, 100% agreed. from what the rumour mill is saying so far, Intel went absolute bonkers, balls to the wall, absolute limit in their manufacturing process with regards to the upcoming 36 core HEDT part. that monolithic silicon is literally the absolute max area size they can fit in their machines to be exposed and etched. even if they were willing to add cost, wattage and area size, it would literally NOT be possible from a technical standpoint to add any more cache or cores to that monster. this basically means Intel has reached the end of the line and i dont think their troubles will get any better if they stick to monolithic designs. AMD can just increase their package and throw in even more chiplets, their only bottleneck being the finetuning of their IF, which they already have down pat pretty much. 

Nvidia will soon be in a similar position, the reason why theyre trying to go so overboard with power and even more focus on RT and DLSS 3.0 means that RDNA3 will likely match or even beat the 4090 in rasterization performance. interesting enough, new rumours also suggest that AMD is currently VERY confident that their SECOND tier part will match the 4090 and their top tier one will beat it in raster, whereas RT theyll catch up but not be able to quite match the 40 series.

 

i can only repeat myself, AMD, love them or hate them, doesnt matter, theyre kicking Intel's & Nvidia's ASS, thus gearing up competition for everyone 🙂 

[Reciever]

8 hours ago, tps3443 said:

 

Honestly, anything HEDT that is new and coming from Intel will be crazy expensive. 
 

I imagine they’ll be competing with AMD Threadripper TRX40 socket right? Which is also crazy expensive. 
 

I highly doubt it’ll be like the X299 days. 

Who knows, not enough info yet

[Etern4l]

2 hours ago, jaybee83 said:

AMD can just increase their package and throw in even more chiplets, their only bottleneck being the finetuning of their IF

 

 

Good point about the benefits of chiplets, but what's IF there?

 

2 hours ago, jaybee83 said:

i can only repeat myself, AMD, love them or hate them, doesnt matter, theyre kicking Intel's & Nvidia's ASS, thus gearing up competition for everyone 🙂 

 

I don't have any particular affinity with Intel, putting aside the fact that I have never owned AMD CPU or GPU 🙂

That said, if it can be shown that AMD can do the job more efficiently than Intel, I'm rolling a Ryzen build.

Part of the problem is the popular synthetic benchmarks. You run CB23 and the only information you get back is a single number which is a proxy of how quickly the CPU would render a small scene. Since I don't do any rendering, I technically don't care, and the amount of memory utilised by CB23 is tiny, which is also inadequate for my purposes. I would need a benchmark focused on numerical computing but also answering the following questions:

* How does the performance scale with the number of logical cores?

* How does the performance fall off with the workload size?

 

Anyone knows of such an off the shelf benchmark?

[Etern4l]

 

A couple of interesting points:

* Nvidia advised him not to use 7950x for the test bench because of compatibility issues... FUD or real?

* The card is basically factory overclocked. Setting power target to 60% yields... 90-95% performance. Insane!

* He noted some issues with FE fans...

 

Otherwise, well: 4090... love at first sight, shame about the price lol

 

Edit: The even more gigantic AIB cards are basically pointless... probably why EVGA got out.

 

 

[Ashtrix]

9 hours ago, Reciever said:

Not that it would help you save any money but isnt Intel HEDT still in the rumor mill?

 

HEDT is dead.

 

Intel and AMD said good bye to the X299 and AMD's X399/sTRX40 as last goodbye to the market.

 

SPR Xeon is delayed to 2023 Q1, that means it will not hit the retail anytime sooner than Q2-3 means more time again for the Xeon based CPUs for consumers. Intel is in similar line with Threadripper Pro,  W series a.k.a Workstation class Xeon W. AMD's Threadripper is dead prematurely, AMD axed all their customer base hanging to the TR CPUs on that sTRX40 boards without a successor. It's all Threadripper PRO series which costs an arm, a leg and a kidney on top. It has more DRAM more Lanes and more everything.. Those things will not be cheap at all. On top the earlier rumors of Genoa EPYC TDP is high because AMD is going all in at 96C/192T monster this round, same for SPR Xeon with 4x14C tiles Intel finally got more cores now at 56C/112T plus Intel's not even using MCM it's EMIB, expensive to manufacture which is why Intel is failing to even get it to their bread and butter Datacenter market.

 

So ultimately HEDT market is gone. It's all W class for Intel XEON W and AMD Threadripper PRO series from now on for that Prosumer market.

 

Look at the  PC Hardware reviews and benchmarks, how many even really actually care about PCIe specifications ? Of mainstream coverage TPU only was the one who covered the PCIe lanes on AM5, Anandtech is another who does it.

 

Again with RTX4090 reviews none of them even mention the Display Port specification on them or PCIe either, this is a full fat real 4K card and nobody even cared to mention it has PCIe 4.0 only, buy a brand new expensive high end PCIe5.0 mobo and have only 4.0 GPU worth $1600 and a DP1.4a spec. CPU and GPU reviews it's all "Gaming" endlessly. So all the PC crowd is also done with those HEDT PCIe lanes high core compute or etc. The market doesn't even exist anymore on top unlike those Haswell E parts these HEDT barely scale any in gaming and have negative impact..

 

Heck only Tom's Hardware did a review on AMD TR PRO. Nobody even batted an eye.

 

Bonus package ? Nvidia killed off SLI NVLink with Ada. HEDT lost even more now,  RTX3090TI FE x2 SLI will be much more than a single GPU RTX4090 for those SLI workloads that scale.. ML/AI and etc esp those which use 48GB of VRAM with SLI.

 

Well it was a good time until it lasted ? Shame how PC is bogged to ash by the Console trends a.k.a  no dev cares about SLI or PC since 100% of the PC games are nowadays only Console ports with added tech some times more optimization and tweaks. And most of the PC buyers are only looking at gaming too.

 

Here's that old video which I'm reposting again, Crysis 3 from 2013 running in SLI with fantastic scaling with 3090 SLI on 8K resolution no DLSS upscaling crap or DLSS fake frames here, what a beautiful sight... maybe the PC exclusive title of the century ? haha. The good ol Crytek which has been dead since a decade now, well we won't be seeing this type of PC only anymore with any company going forward. There's no passion left anymore in AAA space.

 

 

[Papusan]

7 hours ago, Etern4l said:

I mean Intel is in trouble. The 10nm process is really limiting them. The consumer CPUs can barely compete with AMD, then there is the ARM architecture. One thing I would like to see next in their high end consumer offering is at least 2x the amount of cache. This can work wonders even for gaming, as AMD demonstrated.

Yep, Intel is in trouble. But AMD sit in the same boat. 10nm ain't a problem for Intel. It's their love for new Cpu socket every 2nd year. And AMD should be damn glad for this. Otherwise they would loose a lot sales over to Intel. Yep, 10nm isn't what limiting them. It's their stupid MB/socket change for NO reasons.

 

DOWNTURN IN THE PC SECTOR Intel is reportedly planning to cut thousands of jobs

It's not just about adjusting spending on factories and materials, jobs are also to be cut at Intel....

 

AMD Ryzen 7000 is off to a slow start, Zen 4 sales are not good techspot.com

No one is buying the new CPUs

 

I'm sure less sales of new modern processors to help on the Cpu bottleneck due 4000 series graphics cards as 4090 will help Jensen to burn in with his new gen overpriced graphics cards. I don't think Nvidia have understood what times we live in now. 
 

[Etern4l]

18 minutes ago, Papusan said:

Yep, Intel is in trouble. But AMD sit in the same boat. 10nm ain't a problem for Intel. It's their love for new Cpu socket every 2nd year. And AMD should be damn glad for this. Otherwise they would loose a lot sales over to Intel. Yep, 10nm isn't what limiting them. It's their stupid MB/socket change for NO reasons.

 

DOWNTURN IN THE PC SECTOR Intel is reportedly planning to cut thousands of jobs

It's not just about adjusting spending on factories and materials, jobs are also to be cut at Intel....

 

AMD Ryzen 7000 is off to a slow start, Zen 4 sales are not good techspot.com

No one is buying the new CPUs

 

I'm sure less sales of new modern processors to help on the Cpu bottleneck due 4000 series graphics cards as 4090 will help Jensen to get rid of his new gen overpriced graphics cards from the shelves. I don't think Nvidia have understood what times we live in now. 
 

 

Sure, the socket flipping is not great for the consumer, but obviously a 5nm process would have helped Intel? 🙂

Put another way, it's amazing that they can still compete with a 10 nm process. Put yet another way, AMD kind of sucks for not being able to completely blow Intel out of the water at 5nm..

 

Anyway, the way the world is currently going, we probably need toughbooks, solar power generators (no diesel, sorry), and a good supply of slingshots, clubs and maybe some abacuses to keep the science going when power finally goes out....